Montage returns with each mosaic an “area file” in FITS format that contains the total amount of sky area (steradians) that contributed to the value of a pixel. It weights the output flux by the “areas” of the contributing inputs.
The file is a useful quality assurance tool and comes into play in three places:
- Around the edges of a reprojected image, where there may be a pixel where only a tiny area contributes to its value;
- When the projection covers enough sky that the projection distortions onto the image plane return different areas for pixels on the edges versus in the middle; and
- When are co-adding values from multiple images where the previous two effects require different weightings for the contributing pixels.
A concrete example will help illustrate how the values are calculated . Suppose you reproject image A and an edge pixel in the input has a value of 1.523 but only overlaps the corresponding output pixel by 10%.The output pixel flux value would still be 1.523 but
the relative area value would only be 0.1. A second image B covers the same output pixel with a value of 1.621 but completely, so the area is 1.0. Then when we coadd, we want to give the coadded pixel the value of
(1.523*0.1 + 1.621*1.0) / (0.1 + 1.00) = 1.612
i.e., weighting the output flux by the “areas” of the contributing inputs. The “area” of the output pixel (which is now more of a weight for future use) is the combined 1.1. If I have lots of input images for the same output location, I can easily have fairly
Adapted from material prepared by Dr. John Good (IPAC)